Janet D. Conway
DEFINITION
Knee arthrodesis is an excellent salvage option for infected total knee arthroplasty (TKA) and for severe trauma about the knee. It is a durable solution that lasts a lifetime and allows for a stable, painless extremity for ambulation.
The energy expenditure for walking with a knee fusion is less than the energy expenditure needed to walk with an above-the-knee amputation. This is important for elderly patients with associated comorbidities who often are unable to walk with such an amputation.9
In the younger, posttraumatic population, the knee fusion is a durable option that allows for the most physically demanding activities as opposed to above-the-knee amputation or TKA.
Ipsilateral hip arthritis or significant back pain and arthritis are relative contraindications to knee fusion, although for some patients, knee fusion is the best option despite relative contraindications.
ANATOMY
The approach for knee arthrodesis is anterior.
The relevant anatomy depends on the previous surgical procedures that the patient has undergone. In the case of the infected TKA, the extensor mechanism often is no longer present and a soft tissue defect occurs anteriorly.
The best bone contact for knee arthrodesis is achieved by elevating the posterior capsule off the distal femur and proximal tibia.
Directly posterior to the posterior capsule are the popliteal artery, popliteal vein, and sciatic nerve as it branches into the posterior tibial and peroneal nerve. Great care must be taken in the area of these vessels (FIG 1).
When using a long intramedullary rod for knee fusion, the piriformis fossa is an important landmark for the entry of the nail.
At present, trochanteric-starting intramedullary knee fusion rods are not available.
PATHOGENESIS
The conditions necessitating knee arthrodesis include infected TKA and severe trauma about the knee destroying the joint.
Trauma, destruction, or débridement as a result of infection leads to extensive bone loss.
The most common way to fill this bone defect is by acutely shortening the limb.
The average shortening associated with knee arthrodesis after infected TKA is 4 cm, which necessitates a shoe lift to equalize the limb lengths.8
In cases with significant amount of shortening—more than 5 cm—concomitant lengthening can be performed in the proximal femur.
In cases with substantial bone loss (7 to 10 cm) and with soft tissues and vessels that do not allow acute compression, gradual compression of 2 mm/day with an external fixator is an option for knee arthrodesis, although this still leaves the limb significantly shortened.
My preference in such situations is for bone transport—either double or single level—in which the overall length of the limb is maintained and the large bone gap at the knee is filled with the transported bone. This strategy works well in the young, traumatized patient and in the elderly patient with infected TKA.
NATURAL HISTORY
The natural history of these severe traumatic and infectious processes occurring about the knee is poor. Without a stable lower limb, the patients are unable to bear weight.
The alternative to knee fusion is above-the-knee amputation, which allows the young, healthy posttraumatic population to walk with an above-the-knee prosthesis.
Above-the-knee amputation in the elderly population with infected TKA usually results in the patient's becoming nonambulatory.9
The natural history of a knee arthrodesis is extremely durable, lasting the patient's lifetime.3
As the patient ages, however, contralateral hip arthritis (secondary to the longer length of the contralateral limb) and arthritis of the spine (secondary to the increased motion across the spine) can develop. These problems can be addressed separately, if and when they occur.
FIG 1 • Proximity of the posterior neurovascular bundle to the posterior aspects of the femur and tibia at the level of the residual knee joint. Also, the bones are very subcutaneous anteriorly and do not have good blood supply secondary to the relatively avascular quadriceps and patellar tendons and lack of good muscle with good blood flow to the bones anteriorly.
PATIENT HISTORY AND PHYSICAL FINDINGS
The history pertinent to preoperative assessment for knee arthrodesis is a list of all surgical and traumatic occurrences about the knee.
This includes previous flap surgeries for coverage of the anterior knee and integrity of the extensor mechanism.
Other important issues regarding the history include comorbidities, such as peripheral vascular disease, smoking, diabetes, ambulatory status, social resources, and steroid use.
These comorbidities will affect the patient's ability to heal a fusion.
The surgeon's choice of technique should take into account these comorbidities to maximize the patient's chance for a successful outcome.
Physical examination of the hips and ankles is important to assess the integrity of the remaining joints that will be compensating for the resultant lack of knee motion.
Any equinus contracture can easily be addressed at the time of surgery with Achilles tendon lengthening or gastrocsoleus recession.
The examiner should palpate the dorsalis pedis and posterior tibialis pulses. If they are poor, a vascular evaluation is obtained.
The skin of the anterior knee is assessed for scars, previous flaps, defects, and integrity. If the skin condition is poor, the surgeon should consider alternative wound-closure techniques postoperatively, such as a wound vacuum or preoperative plastic surgical consultation.
IMAGING AND OTHER DIAGNOSTIC STUDIES
The most important imaging studies are long standing anteroposterior (AP) view erect lower limbs radiographs and long lateral view radiographs (FIG 2).
These radiographs allow the surgeon to assess the present and predicted limb-length discrepancy after knee fusion.
With defects in the knee region greater than 5 or 6 cm, acute compression intraoperatively might lead to vascular compromise from vessel kinking.2 Therefore, these radiographs are essential for determining the method of fusion.
Any residual bone cement in the medullary canals can also be visualized on these views, allowing the surgeon to plan for the proper tools to remove it at the time of surgery.
Magnetic resonance imaging can be helpful in determining the extent of any infection in the distal femur or proximal tibia. Care must be taken with interpreting the images, however, because bone edema can be misinterpreted as osteomyelitis and result in too aggressive a resection.
DIFFERENTIAL DIAGNOSIS
Indications for knee fusion:
Infected revision TKA
Severe trauma about the knee preventing reconstruction
Reconstruction after tumor resection
NONOPERATIVE MANAGEMENT
Nonoperative management of a knee with a significant bone defect after trauma or joint infection is very difficult. These patients typically have an unstable limb for weight bearing and need a cast or brace for support.
This resection arthroplasty option usually is reserved for nonambulators and patients who are too medically ill to undergo a substantial surgical procedure that involves significant blood loss.
SURGICAL MANAGEMENT
Surgical intervention is the most effective way to obtain a knee arthrodesis.
It requires a patient who is medically stable enough to undergo a 2- to 3-hour procedure with 500 to 800 mL of blood loss.2
FIG 2 • AP view erect lower limbs radiograph and a long lateral view radiograph. This figure also demonstrates how to obtain the radiograph. Valuable information regarding alignment and limb-length discrepancy is determined by using these films. The pelvis is leveled with blocks before the radiograph is obtained to ensure accurate measurements. With the pelvis leveled, the patient cannot compensate for the limb-length discrepancy with equinus of the short limb or knee flexion of the long limb. The radiograph is obtained from 10 feet away by using a 51-inch cassette.
Preoperative Planning
Proper preoperative planning begins with the critical points outlined in the Patient History and Physical Findings and Imaging and Other Diagnostic Studies sections.
Essential to preoperative planning is determination of the resultant gap that will be present at the site of the knee fusion (Table 1).
Acute compression of the gap with intramedullary rod fusion should be reserved for gaps no more than 5 to 6 cm. Acute compression greater than that can cause vessel kinking and ischemia to the lower limb.
Gaps greater than 5 to 6 cm can be managed with gradual compression or bone transport to fill in the defect. Bone transport with a fixator allows the limb to remain at a desired length (1 cm shorter than the contralateral limb) and fills the gap with healthy bone from the proximal tibia or femur.
In cases with large gaps, gradual compression without lengthening, achieved by using an external fixator, will eliminate vessel kinking, but the resultant limb-length discrepancy might be undesirable for the patient and require him or her to wear a 2- to 3-inch shoe lift, which can be very cumbersome for ambulation.
It is vital to discuss the goals of knee arthrodesis with the patient before surgery. The surgeon must be sure that the patient is willing to accept a large shoe lift; if not, the patient must be willing to undergo additional steps to ensure that the limb is of acceptable length.
The strategy for knee fusion is patient dependent, and the lengthening can be performed at a second surgical setting or during the same surgical setting. It is important to realize that with concomitant lengthening, the rate-limiting step in the complete healing process is the fusion site, not the regenerate bone formation.
All knee fusions after infected TKA are bone grafted. This is done when there is no evidence of infection.
For patients undergoing two-stage procedures (infection eradication and spacer plus clean fusion with intramedullary rod or plates), the bone grafting is performed at the time of the fusion.
In patients who have external fixation and for whom fusion is initiated at the time of the infection eradication surgery, bone grafting is performed at a second surgical setting.
Regardless of the method of fusion, fusion after infection requires two surgical procedures.
My preferred method for obtaining a large volume of bone with minimal donor site morbidity is to use the ReamerIrrigator-Aspirator System (Synthes, Inc., Paoli, PA). The reamer is used to obtain a large volume of bone from the contralateral femoral canal through a small (2-cm) incision.
Hardware Considerations
Long Knee Arthrodesis Rods
The long intramedullary rod is an excellent method of fixation for knee arthrodesis. It is well tolerated by patients and provides good neutralization of the forces generated by the muscles around the knee.
Biomet, Inc.: Biomet Trauma (Warsaw, IN), Smith & Nephew (Memphis, TN), and Stryker Orthopaedics (Mahwah, NJ) are three companies that currently offer long knee fusion rods.
The Stryker intramedullary rod is the only one with 5 degrees of valgus to counteract the mild varus mechanical axis of the limb with insertion of a straight rod.
The Stryker rod also has the ability to compress the fusion site after rod insertion and locking.
This is done by a proximal compression bolt end cap inserted in the proximal rod. The compression bolt end cap sinks into the rod and engages the proximal interlocking screw in the dynamic slot. This allows for an additional 1 cm of compression at the fusion site, which can be helpful after rod insertion.
Short Intramedullary Fusion Rods
Short intramedullary fusion rods are rods that couple at the knee arthrodesis site.
The one that is commercially available is called the Wichita Fusion Nail (Stryker).
The device (FIG 3) has two separate segments of different diameters for the femur and tibia. Each segment is fixed with interlocking screws, and a coupling device is then used to engage and compress the bone ends.
The device works very well for the primary knee arthrodesis that has good metaphyseal bone.1
Without good metaphyseal bone and tight-fitting rods in the femoral and tibial canal, the Wichita Fusion Nail does not provide enough stability to neutralize the long lever arms across the knee.
Preoperative planning is essential to ensure that the tibial and femoral rods will have good fit in the bone. The femoral rod is 14 cm long, and the tibial rod is 16 cm long.
If the distal femur and proximal tibia appear as “cortical shells” without good bone, another method should be chosen.
External Fixation
Many systems are available for external fixation, but my preferred method is a biplanar Orthofix LRS external fixator (Verona, Italy; FIG 4).
This system has two long smooth rails (65 and 80 cm) that are mounted anteriorly and laterally from the hip to the ankle to adequately neutralize the long lever arms across the knee. It also allows for purchase of the bone in an area not violated by any long revision TKA stems.
The advantages to external fixation are that at the completion of the fusion, no hardware remains as a nidus for recurrent infection and the external fixator can be applied at the same surgical setting as the removal of infected TKA components or débridement of osteomyelitis. This allows for immediate initiation of bone contact and knee fusion.
FIG 3 • The Wichita Fusion Nail. Several figures from the technique guide depict the separate components, the insertion handle, and the compression mechanism. Note the separate femoral and tibial components with the coupling device. (A–C, Courtesy of Stryker, Mahwah, NJ.)
It is not recommended to perform bone grafting at this setting but to wait until the soft tissue envelope is stable and no fear of infection at the knee fusion site remains. This often is done 6 to 8 weeks after the application of external fixation.
Also, with a circular or biplanar system, external fixation can be set up so that as the fusion is uniting, the resultant limb-length discrepancy can be diminished by performing an osteotomy of the proximal femur or distal tibia at the same surgical setting.
Circular fixation, with the rings sitting at the medial aspect of the thigh, can be cumbersome for large or elderly patients—hence the description of the biplanar Orthofix external fixator.
Any monolateral system can be set up in a biplanar fashion. The system used must span the length of the femur and tibia to achieve rigid fixation.
Plates
Plates are not commonly chosen implants for knee fusion because of the bulkiness of the plates and the lack of soft tissue envelope in the anterior aspect of the knee. However, plates might be preferred for patients with total hip arthroplasty above the desired fusion site.
The ideal construct for plates should be a 90-90 construct in which an anterior plate is used to counteract the flexion–extension forces and a medial or lateral plate is used to counteract the varus–valgus forces.
The plates should initially be used in compression mode, and the remaining screw holes should be inserted in a locking fashion into a locked plate.
FIG 4 • Biplanar mounting of the Orthofix LRS external fixator for fusion of the knee.
FIG 5 • Intraoperative patient positioning used for all knee fusion cases. The entire limb is prepared, with a bump under the ipsilateral buttock. The bump allows for access to the proximal femur for intramedullary rodding or placement of external fixator pins.
Many plating systems are commercially available with locking capability (Synthes, Smith & Nephew).
Positioning
The positioning for knee fusion is the same for each surgical approach.
A bump is placed under the buttock to allow for visualization of the femoral neck and head on the lateral view radiograph.
The entire limb, including the foot, needs to be visualized to ensure proper rotation of the limb during surgery and also to assess the pulse during the procedure (FIG 5).
Approach
The basic approach to knee fusion is anterior.
Without any soft tissue compromise, the best incision for knee fusion is a transverse one that will be easily approximated once the limb is acutely shortened.
A longitudinal incision becomes difficult to close once the bone ends are shortened (FIG 6A,B).
The transverse incision allows for removal of a TKA without difficulty.
If knee fusion is performed as a solution for infection (post-traumatic or infected TKA), the fusion often can be initiated at the same surgical setting as the débridement.
The preferred technique is to use a Mayo stand for the débridement (FIG 6C).
The scrub technician places the required instruments onto the Mayo stand, and the surgeon retrieves them as needed. When débridement is completed, the “dirty” Mayo stand is moved away and the limb is reprepared, additional clean drapes are placed, and new gowns are used.
New Bovie, suction, and light handles are used for the clean procedure.
A high-speed burr with continuous cooling irrigation is used for débridement of clean, healthy, bleeding bone and to obtain good bone surfaces for maximum bone contact during compression.
A Versajet hydroscalpel (Smith & Nephew) is used to achieve thorough débridement of the soft tissue and especially the posterior capsule. Use of the Versajet is a very safe way to achieve maximum soft tissue débridement.
The back of the capsule is always freed from the posterior aspect of the bone to allow good bone contact without compromising the vessels behind the capsule. This usually is done very carefully with the Cobb.
Once the bone ends are prepared, any of the methods described below can be used to stabilize the fusion.
FIG 6 • Transverse anterior knee incision during the surgical procedure (A) and postoperatively (B). Note the excellent exposure and the excellent wound healing postoperatively. C. A separate Mayo stand is used only for the “dirty” or débridement portion of the surgical procedure.
TECHNIQUES
LONG INTRAMEDULLARY ROD INSERTION
Incision and Exposure
This technique begins with a transverse incision at the knee, centered between the tibia and the femur as determined using fluoroscopy.
A transverse incision is designed to heal when the bone ends are compressed.
After the initial exposure, cutting transversely through any remaining patellar tendon or quadriceps tendon, the bone ends are exposed.
All soft tissue that is tethering the bone ends' ability to compress must be released without jeopardizing the blood supply to the fusion site.
This includes, as mentioned above, freeing the capsule from the posterior aspects of the femur and tibia in a very careful fashion with the Cobb to get direct contact of the bone ends.
Bone is resected in a careful fashion to minimize the resultant shortening of the limb and to achieve the maximum bone contact possible. Any necrotic bone must be resected.
Reaming the Tibia and Femur
Once the bone ends are prepared, the tibia is reamed first.
The femur must be reamed to the same diameter that the tibia was reamed so that maximum stability of the fusion can be achieved with the nail.
Overreaming the femur will disallow the best tight fit of the nail into the femoral canal.
This is secondary to the fact that the nail diameter is the same from the proximal femur to the distal tibia.
Previous knee fusion rods had different diameters for the femur and tibia but are no longer commercially available. If this type of rod were available, the femur would be reamed to 1 mm more than the available diameter for the femur (TECH FIG 1).
The tibia and femur are reamed separately over a guide rod.
The tibia is reamed in an antegrade fashion, whereas the femur is reamed in a retrograde fashion.
TECH FIG 1 • Lateral depiction of antegrade reaming of the tibia and then retrograde reaming of the femur. The trochanteric fossa is also known as the piriformis fossa.
The guide rod is tapped out of the proximal femur through the piriformis fossa. This is an extremely easy way to find the nail insertion site at the proximal femur.
Care must be taken, however, to ensure that this does not place the starting point too medial. If this seems to be the case, the proximal starting point is found with a Steinmann pin proximally to ensure that the dreaded complication of femoral neck fracture will not occur from too medial a starting point.
The guide rod that was inserted in a retrograde fashion through the piriformis fossa is then pushed out through the skin through a small stab incision.
The proximal 8 cm of the antegrade knee arthrodesis nail usually is 13 to 14 mm in diameter. The proximal portion of the femur is then reamed antegrade to preserve the tight fit of the nail in the remainder of the femoral canal.
Inserting the Rod
Rod insertion is the most important part of the case in two respects.
The first is to ensure that the bone ends are lined up and there is compression at the fusion site with insertion. It is critical that the bone ends are lined up evenly with the guide rod inserted the entire distance from the femur to the ankle because the rod can still deviate in the soft bone with insertion and violate the cortical wall. Holding the bone ends compressed will also ensure maximum contact at the fusion site after the nail is completely inserted.
Second, ensuring the proper rotational alignment of the limb at this stage is critical. The initial position of the limb when prepared was on a bump to internally rotate the limb. When inserting the rod, the limb is adducted. Because of the internal rotation from the bump, my final position for the foot once the limb is adducted is perpendicular to the floor. This ensures some external rotation of the limb once the bump is removed (TECH FIG 2).
TECH FIG 2 • Position of the limb as the intramedullary rod is inserted at the hip. The critical portion of the procedure is to set the rotation with the foot perpendicular to the floor to obtain some final external rotation when the bump underneath the buttock is removed.
Do not count on fixing rotation once the nail has engaged the tibia. Because of the tight fit of the nail in the tibial canal and the anterior bow of 5 to 7 degrees—and, if a Stryker knee arthrodesis rod is used, an additional 5 degrees of valgus that is in the nail—rotating the tibia once the rod is fully engaged in the tibia can lead to tibial fracture.
Once the rod is inserted and locked proximally with a guide arm, additional compression can be achieved at the knee fusion site by holding the foot and driving the rod in more with three or four more mallet slaps to the insertion handle.
The limb is then taken out of adduction and locked distally with the use of a fluoroscopy-guided “perfect circle” freehand technique.
Make sure that adequate compression is maintained until the locking screws are inserted.
The Stryker knee arthrodesis rod has the added feature of a compression screw that can be inserted proximally and allows up to 1 cm of additional compression after the distal interlocking screws are inserted.
Wound Closure
After nail insertion, the incisions are closed with absorbable monofilament suture.
If nail insertion was performed in a clean fashion, bone graft, bone morphogenic protein, or both can be added to the fusion site before closure. This is especially helpful if minor gapping is present at the fusion site.
SHORT INTRAMEDULLARY ROD INSERTION
The surgical approach can be a standard medial parapatellar approach or through the transverse incision mentioned above.
Once the bone ends are exposed, an intramedullary guide is used to align the distal femoral and proximal tibial cuts in about 5 degrees of flexion and neutral varus–valgus alignment.
A trial reduction of the bone ends is performed after the bone cuts to check the bone apposition and alignment. Great care should be taken to obtain good bone contact at the fusion site.
Another factor to keep in mind is the resultant limb shortening. The average knee arthrodesis is 4 cm short. Too aggressive a resection will result in more shortening and the patient will need to wear a large, bulky, and awkward shoe lift.
Another way to ensure some flexion in the system is to ream the femur from distal-posterior to proximal-anterior. Because the tibial canal is smaller, it is more difficult to ream the tibia in an eccentric fashion like the femur.
Once the femur is reamed, the femoral rod is inserted and locked with the targeting arm. Two screws are placed in a lateral-to-medial fashion.
A slot is then cut into the tibia to allow for the coupling mechanism between the two rods and the tibial guide arm for the interlocking screws.
The bone plug is saved for grafting at the end of the case.
There are two options for the tibial screws. Use the ones that will capture the best bone.
The tibia is then inserted and locked in a medialto-lateral fashion.
When inserting the screws, placing a bump underneath the knee will ensure that the femur and tibia are locked with the rods in some flexion, ideally 5 degrees.
Once the rods are placed and locked, an additional femoral slot can be removed to allow further visualization of the coupling mechanism. This bone plug is also saved for grafting at the end of the case.
Make sure that the rotational alignment is in neutral to 5 degrees external before completely engaging the tibial rod in the femoral rod and screwing down the compression mechanism. Tighten the screw to get good compression at the bone ends. Take care not to overtighten, or the bone ends might fracture.
Once the fusion site is compressed, the bone plugs are replaced as bone graft and the incision is closed.
Full weight bearing is allowed after this procedure if the surgeon is satisfied with the amount of bone contact at the fusion site.
EXTERNAL FIXATION
Application of the Lateral Rail
The first step after adequate exposure of the bone ends and débridement of any residual infection is the application of the lateral rail.
The lateral rail is set up with four clamps: two for the tibia and two for the femur.
The most proximal clamp is placed at the level of the lesser trochanter, perpendicular to the femoral shaft on the AP view and in the midshaft of the femur.
It is very important when placing the proximal femoral pins to ensure that they are not positioned too anteriorly in the femoral shaft. This is a major stress riser and can cause a femoral fracture.
Once one proximal pin is inserted, the most distal tibial pin is then inserted perpendicular to the shaft of the tibia in the AP view.
Rotation of the limb is set with this pin insertion.
I prefer to have the toes perpendicular to the floor because when the proximal bump is removed from under the buttock, the foot will be about 10 degrees externally rotated.
TECH FIG 3 • A,B. Radiographs of a posterior plate in a patient with scarring of the anterior soft tissue envelope. C,D. Ideal positioning for the plates (90-90) and alternative plating positions (ie, medial and lateral). (C,D:Adapted from Conway JD, Mont MA, Bezwada HP. Arthrodesis of the knee. J Bone Joint Surg Am 2004;86A:835–848.)
After the pin insertions, the middle clamps are positioned (TECH FIG 3).
Positioning of the clamps is variable and based on the bone quality at the proximal tibia and distal femur.
Ideally, the greater the span of the clamps, the better fixation of the bone.
A lateral view radiograph is obtained to check the middle clamps.
The clamps often are too posterior to hit the bone and need to be moved proximally or distally accordingly.
The clamps can also be adjusted by adding a half or full “sandwich” to the clamps to raise the pin insertion site more anteriorly. It is preferable to use the sandwiches to raise the pin insertion sites as opposed to moving the clamps further away from the knee joint.
It is at this stage that the flexion of the knee can be “set” into the system.
More flexion will necessitate raising the middle lateral two clamps more anteriorly to hit the bone.
My usual position is 5 degrees of flexion. This minimizes any additional limb shortening from excessive flexion.
In this position, the two middle clamps need one full sandwich to hit the bone.
After the insertion of one pin in each clamp, the remaining pins are inserted for a total of eight half-pins (two pins per clamp).
The preferred half-pins are hydroxyapatite-coated and are inserted so that the thread distance is the same as the diameter of the bone.
If the threads remain outside the bone, the pin is weaker than if the threads were buried to the shank.
Checking Alignment and Mechanical Axis
Once all the pins are inserted, a Bovie cord is used to check the mechanical axis of the limb. The limb is first placed in the “patella forward” position. Under fluoroscopic guidance, the Bovie cord is used as a straight line from the center of the femoral head to the center of the ankle. After confirming these points, fluoroscopy is used to check where this line or mechanical axis lies at the knee. It should be in the center of the knee or slightly medial. If it is not, the tibial pins can be translated in the clamps more medial or more lateral until the mechanical axis is acceptable.
Once this is the case, the pins are secured in the clamps and the tibial clamps are linked with a compression–distraction device. The proximal femoral clamps are secured to the rails.
A second compression–distraction device is then placed between the tibial and femoral clamps and compressed.
The knee fusion site is visualized during the compression to ensure good bone contact and to make sure there is no soft tissue interposition at the bone ends.
Wound Closure
Once the bone ends are opposed and compressed, the anterior knee wound is closed, usually over a drain.
Once the lateral rail is applied and the wound is closed, it is very easy to apply the anterior fixator. The long rail is placed anteriorly with four clamps set up in the same fashion as the lateral rail. The clamps are placed so as not to hit the other pins upon insertion of the anterior pins (TECH FIG 4).
Additional compression can be obtained at the knee fusion site in the office by using the compression–distraction device between the femoral and tibial clamps.
TECH FIG 4 • The mounted Orthofix LRS after the lateral fixator is applied.
PLATING
Plate Size
The number of holes in the plates chosen depends on the bone available for fusion in the tibia and femur.
If a total hip arthroplasty is present, stopping the plate immediately distal to this can be a stress riser. In such cases, sliding the plate a few holes past the total hip arthroplasty stem and using unicortical screws in the region is helpful.
The ideal number of holes is 11: 5 for femoral fixation, 4 for tibial fixation, and 2 left empty at the fusion site.
Exposure
The surgical technique begins with the same exposure as previously mentioned.
A transverse incision can be used, and the plates can be inserted percutaneously in both the anterior and mediolateral plane.
Fluoroscopy is used to ensure that the plates are flush and securely fastened to the bone.
The important step is preparation of the bone ends and good bone contact.
When using plates, the area must first be “sterilized” with the two-stage approach of using an antibioticcoated cement spacer and then 6 weeks of antibiotics.
Once this stage is completed, the plates are inserted as a “clean procedure.” This allows autogenous bone graft to be inserted at the fusion site with bone morphogenic protein.
After preparation of the bone, the alignment is assessed using the Bovie cord test.
Anterior Plate
When good bone contact and good alignment are achieved, the plates are applied. The anterior plate is easier to apply first.
The first step is ensuring that the proximal and distal ends will be well approximated to the bone. This is done with a provisional fixation pin at both ends.
Great care must be taken to ensure that the rotational, sagittal, and coronal alignments are maintained while the plate is applied.
A four-pin temporary lateral fixator can be helpful to achieve alignment and to hold the alignment while the anterior plate is applied. A good assistant can be as helpful as a temporary fixator.
Once the alignment is good and the plate is applied with the provisional pins, the next pins to be inserted are close to the fusion site—one on the femoral side and one on the tibial side—placed in compression mode.
This does two things: compresses the fusion site and pulls the plate down to the bone.
Once the two screws are inserted, the remaining screws can be placed in a locked mode.
This allows for maximum rigidity of the construct so that some weight bearing can be initiated immediately postoperatively.
Medial or Lateral Plate
After anterior plate insertion, the medial or lateral plate can be applied. This is the easier of the two plates to be inserted because the alignment is now rigid.
Medial or lateral is best determined by the amount of soft tissue coverage, with the plate being applied where there is the best chance for the best soft tissue envelope.
Occasionally, a posterior plate can be applied on the lateral side of the knee when the anterior soft tissue is too deficient to cover the plate (TECH FIG 5).
This requires repositioning the patient in a prone position to apply the plate.
TECH FIG 5 • The mounted Orthofix LRS with the completed frame.
STRATEGY FOR SUBSTANTIAL BONE LOSS: TRANSPORT OVER A NAIL
For bone loss of more than 5 cm and noncompressible soft tissue defects at the knee secondary to extensive scarring, bone transport is the best option to fill the defect.
The technique begins with determining the extent of the gap. If the gap at the knee will be more than 10 cm, a double-level transport can be performed.
The first step for transport over a nail is to insert the long intramedullary rod as described earlier.
When inserting the rod, make sure that the limb does not inadvertently lose any length. This can best be accomplished by determining the rod length to be used preoperatively from an erect lower limbs radiograph. The length of the normal side can be used as a reference as long as significant shortening of the affected limb is not also present. The affected limb cannot be acutely lengthened because the soft tissues about the knee are not compliant.
Ideally, the affected limb should be 1 cm short to allow clearance of the foot when ambulating. The average knee fusion shortening is 4 cm, and anything up to this amount is tolerable. Any limb shortening more than this can be addressed with the lengthening over a nail technique at the completion of the transport.
The rod diameter chosen for the transport is 10 mm. This allows the transport segment to slide over the rod when the canal is reamed to 12 mm.
Determine the segment to be transported.
The femur is the preferred segment because of the need to perform only one osteotomy and because of the detrimental effects that proximal tibial transport can have on the ankle (equinus).
If, because of the large segmental defect, tibial transport proximally is necessary, the fibula should also be osteotomized at the midshaft and a distal syndesmotic screw should be placed to prevent any proximal fibular migration.
Mark out the osteotomy site of the transported segment.
Once the guide rod is inserted into the femur and tibia, the rod is backed out past the level of the osteotomy and the osteotomy is predrilled with multiple drill holes before reaming.
This allows the reamings to exit out the osteotomy site and to “bone graft” the regenerate site.
The first step is reaming the intramedullary canal of the tibia and femur to 12 mm.
This can be done through the knee, reaming the tibia and femur separately, or from the hip using long 80-cm reamers (Biomet Trauma, Stryker).
Once the rod is inserted and locked at the desired length, the monolateral external fixator is applied.
Applying the monolateral frame to move the transported segment over the nail requires inserting the pins so that there is no contact between the rod and the pins.
With this technique, because the rod and pins are so close, there is a 5% chance for infection of the rod.6
My preference for the femur is to use a lateral Orthofix LRS frame with two pin clamps. Three halfpins are inserted into the proximal clamp, and three half-pins are inserted into the distal clamp. The pins are inserted by using a cannulated wire technique.
The cannulated wire technique starts with a 1.8-mm wire inserted perpendicular to the rod on the AP view fluoroscopic projection but away from the rod by a few millimeters on the lateral view projection.
The most common location in the femur for these pins is proximally and posteriorly at the level of the lesser trochanter (TECH FIG 6).
Once the wire is inserted, it is confirmed using fluoroscopy. The fluoroscopic view must show the wire “on end.”
This is to confirm that when the pin is drilled and placed, it will not be touching the rod. This technique can be very time-consuming.
Fluoroscopy must be used frequently to confirm that the pins are placed away from the rod.
TECH FIG 6 • Steps involved in the transport over a nail technique. Bone graft and a plate are applied to the docking site and the fixator is removed at the final surgical setting.
Once the wire is in a satisfactory position on the AP and lateral view projections, a 4.8-mm cannulated drill bit is used to drill the near cortex. This drill bit and the wire are then removed, and a solid 4.8-mm drill bit is used to complete the tract for the pin.
Drilling with the cannulated drill bit and then the solid drill bit is important because the cannulated drill bit is not end-cutting and sharp enough to go through the cortical bone of the far cortex. Often, these pins are placed entirely in the cortical bone.
When using the drill, it is imperative that the drill bit not heat up and cause osteonecrosis of the bone.
If this happens, the pin will become infected and a ring sequestrum will develop. Also, an infected pin places the intramedullary rod at risk for contamination.
To prevent this, the drill bit is removed at regular intervals while drilling to be cooled and cleaned with a wet, cool laparotomy sponge.
Once the bone is drilled, a 6-mm hydroxyapatite-coated pin is inserted.
After insertion of the pins with use of the Orthofix clamp as a guide, the frame is removed and the bone is cut with an osteotome.
A small incision is used laterally at the level of the femur.
Often, the bone cannot be completely cut through one incision around the rod. A second incision is then placed anteriorly to complete the osteotomy along the medial femur.
If the tibia is chosen, the incisions are placed anteriorly and medially to obtain access to the lateral cortex and posteromedial cortex, respectively.
Once the bone is cut, the pins are used to carefully rotate the bone and determine that the osteotomy is complete.
When the osteotomy is complete, the fixator is reapplied and the osteotomy site is distracted to ensure that the bone ends will separate.
This is confirmed by using fluoroscopy, and the osteotomy site is then reapproximated.
This completes the procedure after the wound closures and dressings.
Postoperatively, the pins are cleaned daily with saline and redressed with a Kerlix dressing wrapped tightly around each set of pins.
The dressing prevents skin pistoning around the pins and limits the soft tissue trauma that leads to pin tract infections.
Touch-down weight bearing for balance only is permitted postoperatively.
Full weight bearing is permitted once two cortices are present at the regenerate site on the radiographs, once the consolidation phase of bone healing has begun.
Distraction is begun at postoperative day 5 and is continued until the gap is closed at the knee region.
When the gap has closed, the patient is brought back to the operating room for insertion of bone graft at the docking site and percutaneous locked plating at the docking site. The locked plating is essential to prevent the transported bone end from migrating.
Custom rods with predrilled holes to lock the transported segment significantly weaken the rod and are not recommended.
Once the bone graft and locked plate are inserted, the external fixator is removed.
If the limb is still significantly short after the docking of the transported segment, the distal interlocking screws are removed from the rod and the external fixator is left in place to continue lengthening.
Once the desired length is achieved, the patient is returned to the operating room for the insertion of the locking screws and removal of the external fixator.
The patient is allowed full weight bearing once two of four cortices are present on the radiographs.
POSTOPERATIVE CARE
Postoperatively, regardless of the technique, the patient is encouraged to strengthen the hip and ankle.
For patients with external fixators, pin tract infections are likely to occur and are initially treated with orally administered antibiotics.
All patients are given a prescription for an antibiotic to be taken orally, most commonly cephalexin, before discharge and are instructed to start the antibiotic at the first sign of redness, increased tenderness at the pin site, or drainage.
Follow-up office visits are every 2 weeks for patients who are undergoing bone transport or lengthening. Once the consolidation phase starts, only monthly follow-up visits are required.
For patients with external fixation, once the bone has consolidated, the frame is then dynamized in the office 1 month before removal.
Dynamization of the frame usually is performed by taking the tension off the compression–distraction devices. This allows the bone to accommodate more load and become stronger before the frame is completely removed.
If the bone is strong enough, no pain should occur with dynamization in the office setting.
For patients with external fixation, if bone graft was not performed at the time of fusion, a second-stage bone grafting procedure can be performed once there is no longer any evidence of infection, approximately 6 to 8 weeks after the index procedure.
Most patients will need shoe lifts added to the outside of the shoe during the postoperative period.
OUTCOMES
Harris et al5 compared the function of knee arthrodesis with that of constrained TKA and found that the knee arthrodesis patients had better stability and performed more physically demanding activities.
Rud and Jensen11 examined 23 knee athrodesis patients and found that 18 had returned to work.
Most patients should expect to have difficulty with stairs, rugs, and ladders,12 and patients who performed strenuous work before the arthrodesis rarely resume that strenuous work postoperatively.
Rand et al10 reported that seven patients with knee arthrodeses could walk one to three blocks and nine successful knee arthrodesis patients were able to walk more than six blocks.
Compared with the alternative—above-the-knee amputation—knee fusion offers a stable, painless, and uninfected limb for weight bearing. Most knee arthrodesis patients are ambulators, whereas, according to Pring et al9, of 23 patients who underwent an above-the-knee amputation for infected TKA, only 7 were ambulators.
The best way to achieve the best outcomes for these patients with difficult problems is to be thorough in the preoperative discussions regarding what knee arthrodesis can achieve for them.
Realistic patient expectations are critical in achieving successful outcomes.
Although revision TKA might be the more attractive alternative, many patients are not proper candidates for that procedure secondary to a poor soft tissue envelope, bone loss, and recurrent infection.
Hanssen et al4 documented that 50% of patients with infected revision TKA eventually went on to knee arthrodesis.
COMPLICATIONS
The complications associated with knee arthrodesis are related to the increased stress placed on the hip, back, and ankle. Osteoarthritis can occur in those areas.
Takedown of the knee fusion in these circumstances is not recommended, secondary to the extensive complications reported in the literature.7
Other complications that occur include recurrent infection and nonunion.
These complications can be very difficult to treat, considering the many medical comorbidities in the older population.
REFERENCES
1. Christie MJ, DeBoer DK, McQueen DA, et al. Salvage procedures for failed total knee arthroplasty. J Bone Joint Surg Am 2003:85A(Suppl 1):S58–S62.
2. Enneking WF, Shirley PD. Resection-arthrodesis for malignant and potentially malignant lesions about the knee using an intramedullary rod and local bone grafts. J Bone Joint Surg Am 1977; 59A:223–236.
3. Conway JD, Mont MA, Bezwada HP. Arthrodesis of the knee. J Bone Joint Surg Am 2004;86A:835–848.
4. Hanssen AD, Trousdale RT, Osmon DR. Patient outcome with reinfection following reimplantation for the infected total knee arthroplasty. Clin Orthop Relat Res 1995;321:55–67.
5. Harris IE, Leff AR, Gitelis S, Simon MA. Function after amputation, arthrodesis or arthroplasty for tumors about the knee. J Bone Joint Surg Am 1990;72A:1477–1485.
6. Herzenberg JE, Paley D. Femoral lengthening over nails (LON). Tech Orthop 1997;12:240–249.
7. Kim YH, Kim JS, Cho SH. Total knee arthroplasty after spontaneous osseous ankylosis and takedown of formal knee fusion. J Arthroplasty 2000;15:453–460.
8. Oostenbroek HJ, van Roermund PM. Arthrodesis of the knee after an infected total knee arthroplasty using the Ilizarov method. J Bone Joint Surg Br 2001;83B:50–54.
9. Pring DJ, Marks L, Angel JC. Mobility after amputation for failed knee replacement. J Bone Joint Surg Br 1988;70B:770–771.
10. Rand JA, Bryan RS, Chao EY. Failed total knee arthroplasty treated by arthrodesis of the knee using the Ace-Fischer apparatus. J Bone Joint Surg Am 1987;69A:39–45.
11. Rud B, Jensen UH. Function after arthrodesis of the knee. Acta Orthop Scand 1985;56:337–339.
12. Siller TN, Hadjipavlou A. Knee arthrodesis: long-term results. Can J Surg 1976;19:217–219.